/0.5 · Fleets

Put the model to work.

Fleets are workflows and AI agents that run on the Observatory model: the Labelizer, enrichment, the product advisor and AI search. What they do is agreed per customer.

What it is

  1. 01

    The Labelizer labels each product family by ad return (ROAS) and stock return (GMROI): Scale, Keep, Watch, Fix or Dead.

  2. 02

    Every value an agent produces is recorded as produced by an AI agent, with source and time.

  3. 03

    The Labelizer, the product advisor, AI search and enrichment run in production on customer data.

Capabilities

Workflows and agents built per customer scope.

The Labelizer

Labels each product family by ROAS from your ad accounts and GMROI from stock and margin: Scale, Keep, Watch, Fix or Dead.

Product advisor

An agent on your product pages that answers shoppers from the product model.

AI search

Search on your webshop that reads the same product model.

Built per scope

Workflows and agents are agreed per customer, and run on the Observatory model.

In practice

What Fleets did for customers.

Customer result

Stock → cash

Dead stock found and moved

The question
“Which product families tie up stock and ad budget without paying back?”
What the platform did
The Labelizer labels each product family by GMROI and ROAS, so stock that stopped moving stands out and can go into online auctions and targeted campaigns.
Result
Stock that stopped moving, identified and sold through online auctions, with targeted campaigns running now to free up cash.

Owner-reported customer result

Customer result

+0.5 pp

Site-wide conversion rate

The question
“How do we convert more of the visitors we already pay for?”
What the platform did
Enrichment completed the product data, and product advisor agents answered shoppers on the product pages.
Result
From properly enriched product data and product advisor agents on the product pages.

Owner-reported customer result

Illustrative example

Scale

Ad budget that follows profit

The question
“Where should the next euro of ad budget go?”
What the platform did
Product families with a strong ROAS and GMROI are labelled Scale, so campaigns can favour them. Families labelled Fix or Dead get attention first.
Result
One label per product family, from your own data.

Illustrative example

How it works

  1. 01

    Scope

    We agree which workflows and agents you need.

  2. 02

    Build

    We build them on the Observatory model.

  3. 03

    Run

    Agents run on live data; every value they produce is recorded with its source and time.

Technical notes

For your IT and data team.

Provenance
Values an agent produces are recorded as produced by an AI agent, with source and time.
Access
Agents change your systems only through agreed Syncback.
MCP
Agents read the model through the MCP server.
AI models
Works with Claude, ChatGPT and your own agents, or ours.
Sources
Push destinations depend on agreed scope.

Questions

Frequently asked about Fleets.

What is the Labelizer?

A Dataloader agent that labels each product family by ad return (ROAS) and stock return (GMROI). Each family gets one label: Scale, Keep, Watch, Fix or Dead.

Which agents run today?

The Labelizer, the product advisor, AI search and enrichment run in production on customer data.

How do I know what an agent did?

Every value carries who produced it: a person, a process or an AI agent, with source and time.

Can agents change data in our systems?

Only through Syncback, and the writeback scope is agreed per system.

Can we use our own agents?

Yes. Your own agents read the same model through the MCP server.

See what your data can answer.

Talk to us